Association Between Infection With Helicobacter pylori and Chlamydia pneumoniae and Risk of Ischemic Stroke Subtypes
Results From a Population-Based Case-Control Study
Background and Purpose— Helicobacter pylori and Chlamydia pneumoniae have been associated epidemiologically and pathogenetically with coronary atherosclerosis. However, population-based data on chronic infection and stroke are lacking. Therefore, we investigated the association of both bacterial pathogens and ischemic stroke subtypes in a population-based case-control study.
Methods— Patients with first ischemic stroke in the population-based Erlangen Stroke Project were collected as cases. Neighborhood controls were drawn from the study population, matched for age, sex, and place of residence. IgG antibodies to H pylori were measured by enzyme immunoassay, and IgG antibodies to C pneumoniae were measured by microimmunofluorescence technique. Conditional logistic regression was used. Analyses were stratified for etiologic stroke subtypes according to Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria.
Results— A total of 145 case and 260 control subjects were included. Chronic H pylori infection was associated with a higher risk of stroke caused by small-artery occlusion (adjusted odds ratio, 3.31; 95% CI, 1.15 to 9.56) and a lower risk of cardioembolic stroke (adjusted odds ratio, 0.21; 95% CI, 0.06 to 0.71). Overall, elevated H pylori as well as elevated C pneumoniae antibodies were not associated with ischemic stroke.
Conclusions— Our population-based study does not provide evidence of any strong association between the immune response to C pneumoniae as a marker of prior infection and ischemic stroke. Further studies are required to reveal the role of chronic H pylori infection as an independent risk factor for the subgroup small-artery occlusion.
Helicobacter pylori and Chlamydia pneumoniae are notorious for causing chronic infections and have been seroepidemiologically linked to coronary heart disease and atherosclerosis.1,2 In addition, C pneumoniae has been directly recovered from atheromatous plaques.3
Ischemic stroke is a heterogeneous mixture of different stroke subtypes caused by atherosclerotic as well as nonatherosclerotic mechanisms.4 Although stroke is thus pathogenetically related to coronary atherosclerosis, data on chronic infection in cerebrovascular disease are limited.5–10 For a valid and reliable evaluation of the risk factor chronic infection in cerebrovascular disease, the underlying pathomechanisms of ischemic stroke must be considered by stratifying the analyses for different etiologic subtypes. Because direct detection methods for H pylori and C pneumoniae depend on cerebrovascular wall samples retrieved in vivo, which are obviously not available, surrogate markers such as antibody levels must be used to describe the association between stroke and previous infection in an indirect manner.
Therefore, we have investigated, in a population-based case-control sample, whether serological evidence of H pylori or C pneumoniae infection is an independent risk factor in different etiologic subtypes of ischemic stroke.
Subjects and Methods
The Erlangen Stroke Project (ESPro) is an ongoing, prospective, population-based stroke registry in Germany, monitoring a study population of 100 330 inhabitants (December 31, 1997). The methods for case detection and the characteristic of the study area were described in detail elsewhere.11
Selection of Cases
First-ever stroke patients in the ESPro were enrolled in the study if they met the following criteria: (1) first ischemic stroke, (2) stroke treatment in a hospital, and (3) admission to a hospital within the first 72 hours after stroke onset. Stroke was defined according to World Health Organization criteria.12 Cerebral infarction was diagnosed on the basis of results of the first CT scan. Between August 1, 1998, and September 30, 1999, a total of 169 patients with first ischemic stroke were registered in the ESPro: 145 (86%) of them could be included in the present study, 2 (1%) were not hospitalized, 7 (4%) were hospitalized later than 72 hours after stroke onset, 9 (5%) refused, and 6 (4%) died before a blood sample was taken.
Selection of Controls
Neighborhood controls were recruited from the same study population that gave rise to the cases. They were “pair matched” to cases for age (5-year bands), sex, and place of residence (matched by blocks within the study region). The area of Erlangen is subdivided into 112 blocks an average of 0.4 to 1.0 km in size, which define social, demographic, and cultural similarities.
Within each block that formed the place of residence for a case, 5 potential controls were identified from the Registration Office by the Department of Statistics. If <5 potential controls were registered in the same block, the remaining number of age- and sex-matched controls was identified randomly from a block next to the patient’s block.
To improve the response rate of the controls, aims and purposes of the study were published in local newspapers and on television in cooperation with the Regional Community Health Authority. Invitation letters to 3 of 5 randomly drawn controls were sent for every case subject; when no 1:2 ratio could be achieved in a case-control stratum within 4 weeks after the invitation, the 2 remaining controls were invited to attend the study. A total of 676 invitation letters were sent. Eleven of the potential controls had moved with unknown address, 275 sent back their informed signed consent to participate, 12 could not attend the hospital for assessment because of severe illness, and 3 were excluded from the study because of a positive history of preexisting cerebrovascular disease (transient ischemic attack or stroke). A total of 260 controls could be included in the study, yielding a participation rate of 40%. In men a complete 1:2 case-control ratio was achieved. For 30 female cases, only 1 control subject could be recruited from the study population because of the low response rate of older women, particularly among those older than 75 years.
Index Evaluation of Cases and Controls
Data were collected through a face-to-face interview by trained health professionals, including a physical and neurological examination, a detailed review of all available medical records, and fasting blood samples for laboratory analyses. Standardized questionnaires were used for data collection, developed for the European BIOMED program.13 Cases were interviewed and evaluated on day 7 after onset of symptoms; blood samples were taken within the first 24 hours after admission. A total of 82% of all blood samples were drawn within the first 48 hours after stroke onset. Information was obtained directly from the patient or, if the participant was unable to communicate, from a close relative. Control subjects were interviewed and evaluated in the same manner as cases in the Stroke Prevention Clinic of the Department of Neurology between October 1999 and April 2000.
Information on the following potential confounders was assessed: hypertension (reported systolic blood pressure ≥160 mm Hg, reported diastolic blood pressure ≥95 mm Hg, patient’s self-report of hypertension, or use of antihypertensive drugs); diabetes mellitus (fasting blood glucose level ≥120 mg/dL, patient’s self-report of diabetes, or use of antidiabetic drugs); hypercholesterolemia (plasma total cholesterol level ≥240 mg/dL or use of lipid-lowering medication); smoking habits (current smoker, yes or no); and level of education (highest grade completed, dichotomized into high school graduation, yes or no).
Since chronic inflammation might be associated with a higher risk of vascular events,14 statistical analysis was adjusted to it. As an indicator of chronic inflammation, C-reactive protein levels >3 mg/L were defined15 by use of a high-sensitivity assay (N High Sensitivity CRP, Dade Behring Marburg GmbH).
Ischemic stroke was classified according to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria4 into large-artery atherosclerosis, cardioembolism, small-artery occlusion, other determined etiology, and undefined cause. The TOAST classification was performed by 2 neurologists on the basis of a clinical examination as well as the results of standardized diagnostic tests including CT scan or MRI, vascular imaging, and ECG or echocardiography. The overall interrater reliability for the classification was good (unweighted kappa=0.65). Because of small numbers (n=2), subgroup analyses were performed without the stroke subtype of other determined etiology.
Blood samples were labeled with a serial number immediately after blood was taken. Thus, all investigators were blind to case-control status during the laboratory tests. Samples of case and control subjects were prepared, centrifuged, and stored in 2-mL aliquots within the first 2 hours after the blood sample was drawn. Sera were kept frozen at −80°C until the analyses. Immunoglobulin G (IgG) antibody titers to C pneumoniae were analyzed with a microimmunofluorescence test (MIKRO-IFT, Labsystems OY). The resulting antibody titers were classified into titers <1:32, 1:32, 1:64, 1:128, 1:256, 1:512, and 1:1024 on the basis of their serial dilution. A priori, we agreed on classifying C pneumoniae antibodies ≤1:32 as negative or low positive. High-positive titers (≥1:64) were used as cutoff because they were rated most appropriate for seroepidemiological purposes.16 To investigate the reliability of the microimmunofluorescence test, the analyses were repeated in 45 randomly selected specimens. The agreement of the classification was good (quadratic weighted kappa=0.70). IgG antibodies to H pylori were analyzed with a commercial enzyme immunoassay (Pyloriset, BAG). The manufacturer’s recommended cutoff was chosen to indicate H pylori seropositivity. To study the influence of both infectious agents together on an individual’s risk of stroke,17 we further classified cases and controls as having 0, 1, or 2 positive IgG antibody titers to H pylori and/or C pneumoniae. Because of the small number of subjects with no serological evidence of any infection (13 case and 21 control subjects), data from these participants and from participants with 1 positive IgG antibody titer were analyzed together.
Statistical analyses were performed with the SAS 8.0 software packet (SAS Institute). The t test was used to test differences in continuous variables, and the χ2 test was used to test differences in proportions. The sample size of the study was powered to detect an odds ratio (OR) of ≥2-fold with 80% power at a 5% level of significance for all ischemic strokes, on the basis of previous findings.5,6,10,16 To estimate the OR and the resulting 95% CI for the matched case-control pairs, conditional logistic regression was performed. The OR was calculated before and after adjustment for potential confounders. Possible interaction between elevated antibody titers and the other covariates in the model was controlled by adding terms of interaction to the regression model and testing the statistical significance of the resulting coefficient. Subgroup analyses were performed by running different regression models for each stroke subtype. All tests were 2-tailed; statistical significance was determined at an α level of 0.05.
The design of the study was approved by the Ethics Committee (No. 1041/1998) and the Data Protection Committee of the University of Erlangen-Nuremberg. All case and control subjects gave their written informed consent to take part in the study. The participants who were not able to consent for themselves because of severe illness or unconsciousness had given consent through their next of kin or guardian.
The mean age of the 145 ischemic stroke cases was 74.6 years (SD ±10.4), and 77 (53%) were females. Characteristics of case and control subjects are demonstrated in Table 1. Table 2 shows the prevalence rates of seropositivity to H pylori, C pneumoniae, and a concomitant infection with H pylori and C pneumoniae.
In conditional logistic regression (Table 3), elevated IgG antibodies to H pylori were significantly associated with a higher risk of stroke due to small-artery occlusion (adjusted OR, 3.31; 95% CI, 1.15 to 9.56). H pylori seropositivity demonstrated a positive but nonsignificant risk of stroke due to large-artery atherosclerosis (adjusted OR, 6.18; 95% CI, 0.48 to 80.42) on the basis of the small number of cases and controls in this subtype. An inverse, significant relation was found between elevated H pylori IgG antibodies and cardioembolic stroke (adjusted OR, 0.21; 95% CI, 0.06 to 0.71). Because of these differential associations in stroke subtypes, chronic H pylori infection was overall not associated with ischemic stroke (all subtypes combined). Tests for interaction between H pylori seropositivity and the other examined risk factors revealed no significant terms of interaction.
For elevated C pneumoniae antibodies, no significant association with subtypes of ischemic stroke was observed. The risk for first ischemic stroke (all subtypes combined) was also not associated with prior C pneumoniae infection (Table 3). No significant interaction could be found for C pneumoniae antibodies and risk factors. As a secondary analysis, a possible association between C pneumoniae antibodies and risk for ischemic stroke was evaluated for higher cutoff titers (≥1:128; ≥1:256; ≥1:512). Varying the titer cutoffs also resulted in no significant association between elevated C pneumoniae antibodies and ischemic stroke (data not shown).
A simultaneous immune response to H pylori and C pneumoniae was associated with an increased risk of stroke due to small-artery occlusion (adjusted OR, 2.88; 95% CI, 1.03 to 8.09) and a decreased risk for cardioembolic stroke (adjusted OR, 0.28; 95% CI, 0.08 to 0.91).
In response to the possibility that controlling for elevated C-reactive protein levels might result in an overadjustment for chronic inflammatory response, all multiple analyses were repeated without adjustment for C-reactive protein. Removal of C-reactive protein from the model did not alter any of the results considerably.
In our population-based study, IgG antibody responses in H pylori and in parallel infection with H pylori and C pneumoniae were associated with an increased risk of stroke in the subtype of small-artery occlusion and a decreased risk of stroke caused by cardioembolism. However, no evidence was found for any strong association between elevated IgG antibodies to H pylori and C pneumoniae and the overall risk of stroke.
H pylori Infection and Stroke Risk
Data on chronic infection and stroke are limited. Our findings were consistent with the results of the only study that addressed the relation between chronic H pylori infection and ischemic stroke. Markus and Mendall10 reported elevated IgG antibodies to H pylori in the subgroup of lacunar stroke, which is comparable to the subtype of small-artery occlusion.4 The adjusted OR of 2.51 (95% CI, 1.19 to 5.28) was compatible with our findings (OR, 3.31; 95% CI, 1.15 to 9.56). The classic lacunar hypothesis postulated a lipohyalinosis of special arteries, caused by diabetes mellitus and hypertension, as underlying pathogenesis of this subtype.18 In recent years evidence has occurred that lacunar infarcts share also pathomechanisms with atherosclerotic disease.19 A similar trend was demonstrated for stroke caused by large-artery atherosclerosis in the study of Markus and Mendall10 and our study (adjusted OR, 2.17; 95% CI, 1.11 to 4.21 and adjusted OR, 6.18; 95% CI, 0.48 to 80.42, respectively), even if no statistical significance could be achieved in our analyses because of the small number of cases in this stroke subtype. Comparisons between both studies about the role of H pylori in cardioembolic stroke could not be performed because the results of Markus and Mendall10 were reported for the combined subtypes of cardioembolism and undefined cause. In our study H pylori infection was found to be associated with a 0.2-fold decreased risk of stroke by cardioembolism. Cardioembolic strokes are mainly caused by cardiac disorders, such as atrial fibrillation, that lead to a thromboembolic occlusion of cerebral arteries.4 Therefore, our findings are consistent with a positive association of chronic H pylori infection and ischemic stroke caused by an atherosclerotic mechanism of cerebral arteries. However, consideration of stroke as a nonhomogeneous condition and analysis of etiologic stroke subtypes separately result in small numbers among specific stroke subtypes. Therefore, the nonsignificant results in the subgroup of large-artery atherosclerosis need further investigations with larger sample sizes. Furthermore, it is possible that the results for H pylori were attributed to residual confounding, particularly by factors related to socioeconomic status, although we tried to control for them. This might be of particular interest in regard to consideration of H pylori infection as a marker of lower socioeconomic status,20 which seems to be an independent risk factor of vascular disease.21 Childhood social class, which was not regarded in the present analyses, might also be an important confounder, since adjustment for markers of childhood socioeconomic status has recently been shown to weaken a potential association between C pneumoniae infection and coronary heart disease.22
C pneumoniae Infection and Stroke Risk
Five previous studies reported on the association between C pneumoniae infection and stroke risk. All of them,5–7,9 except 1,8 found a positive association between elevated C pneumoniae antibodies and stroke, varying from an OR of 1.71 (95% CI, 1.08 to 2.7)5 to 8.58 (95% CI, 1.07 to 68.82).7 However, a number of methodological aspects must be considered when these findings are interpreted. Previous studies analyzed ischemic strokes and transient ischemic attacks together,5,6 as well as first-ever and recurrent strokes.6 Some of them reported all strokes, including intracranial hemorrhages, subarachnoid hemorrhages, and undefined types,6,7 and 1 study excluded ischemic strokes of undefined cause.9 None of the former studies stratified for etiologic subtypes of ischemic stroke. Different case definitions, combination of incident and recurrent strokes, and analysis of all stroke subtypes together might influence a potential association in both directions. Additionally, most of the findings were based on case-control studies with small sample size,5,6,8,9 which might be prone to selection bias. Two studies chose hospital-based controls,5,6 which may not be representative of the study population or may lead to an overmatching of risk factors.23 One study used population-based controls, selected by random-digit dialing.9 No information was provided about the proportion of the study population with access to a telephone line, which might have caused an underestimation of lower social class in the control group. Differences were also found in measurement methods and definition of C pneumoniae infection. Even if all studies applied the microimmunofluorescence technique to diagnose chronic C pneumoniae infection, divergent antibody titers were used for defining seropositivity, ranging from IgA and IgG ≥1:16 as a cutoff titer9 to IgA ≥1:64 and IgG ≥1:512 as a combined cutoff point.7 If no “reference standard” for seropositivity to C pneumoniae is determined, varying cutoff points might be chosen after data exploration to report a maximal positive association between exposure and disease. Although the results of a microimmunofluorescence assay depend on the raters,24 no previous study checked interobserver agreement of the classification, and 1 did not report on blinding case-control status.9 A positive association between C pneumoniae antibodies and stroke was found only for elevated IgA5,9 or a combination of elevated IgA and IgG antibody titers.6,7 However, a C pneumoniae workshop group recently considered the actual data, which reported elevated IgA antibodies to C pneumoniae to be a marker of chronic infection, to be unconvincing.24 Thus, there is evidence that some of the positive associations between C pneumoniae seroprevalence and stroke risk reported in previous studies might be explained by selection bias, information bias, or residual confounding.
The presence of C pneumoniae in atherosclerotic plaques has been demonstrated in noncerebral vessels, and chronic infection and inflammation are becoming integral parts of the hypothesis of atherosclerosis.2 However, IgG as a serological marker for chronic C pneumoniae infection was not found to have any association with ischemic stroke in our study. In addition to methodological reasons, this result could be due to a different age distribution in our study. The mean age of case subjects was 75 years compared with 36.5 years5 to 68.5 years9 in previous studies. The prevalence of seropositivity to C pneumoniae increases with age.25 Therefore, the involvement of older cases might weaken a moderately positive association between C pneumoniae infection and stroke by higher prevalence rates in cases and controls. Our study was powered to detect an overall association of ≥2-fold between C pneumoniae seropositivity and ischemic stroke. Therefore, the number of case-control subjects within the subgroups was too small to detect a possible moderate association between elevated C pneumoniae antibodies and different etiologic stroke subtypes with a sufficient power. The low participation rate among control subjects might be caused by the older age distribution of our study. Potential controls from the general population within older age groups are more likely to be disabled or suffering from severe comorbidities. Therefore, they might be less willing to participate in any voluntary study than younger subjects, especially if they were not evaluated in their domestic surroundings. Thus, we cannot exclude that some of our findings might be caused by a potential selection of control subjects. As a result of the case-control design of our study, blood specimens were obtained from the cases during the acute hospitalization. Because the infection status of case subjects before stroke remains unknown, we cannot exclude that the disease condition itself might influence the antibody levels. However, we tried to minimize the possibility of changes in serology after stroke by restricting the study population to cases that were admitted within the first 72 hours after onset and by collecting the blood samples within the first 24 hours after admission. Additionally, it is known that serology correlates poorly with the presence of C pneumoniae in vascular specimens.24,26 Therefore, more reliable markers for the diagnosis of chronic C pneumoniae infections are required in epidemiological studies, for example, detection of infected peripheral blood mononuclear cells,24,27 before a potential causal relationship between C pneumoniae infection and stroke can be ascertained.
In summary, our study indicates no strong association between antibody response to C pneumoniae and first ischemic stroke in a population-based sample. This finding is in accordance with recently published epidemiological studies and meta-analyses on coronary artery disease.22,28–30 Further studies are required, however, to confirm whether chronic H pylori infection is an independent risk factor for strokes caused by an atherosclerotic mechanism.
This study was supported by a research grant from the University of Erlangen-Nuremberg (ELAN-Fond 98.08.14.1). The Erlangen Stroke Project is supported by the German Federal Ministry of Health (BMG) for data collection and registration (317-123002/17 00), the German Federal Ministry of Research (BMBF) for research work (N01 EG 9706/5), and the Bavarian Ministry of Health and Social Affairs (VIII 4/9228/24/95).
We wish to thank all participants of the study for their help, especially Drs Hess and Schuster from the Department of Statistics, Drs Lederer and Lohwasser from the Regional Community Health Authority, Dr Gieffers from the Institute for Medical Microbiology and Hygiene, Medical University of Luebeck, Drs Schuettler and von Kegler for investigating the control subjects, and Dr Berger from the Institute of Epidemiology and Social Medicine, University of Muenster, for his helpful comments.
- Received May 3, 2001.
- Revision received June 13, 2001.
- Accepted June 27, 2001.
Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EEIII. Classification of subtype of acute ischemic stroke: definitions for use in a multicenter clinical trial: TOAST: Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993; 24: 35–41.
Wimmer ML, Sandmann-Strupp R, Saikku P, Haberl RL. Association of chlamydial infection with cerebrovascular disease. Stroke. 1996; 27: 2207–2210.
Cook PJ, Honeybourne D, Lip GY, Beevers DG, Wise R, Davies P. Chlamydia pneumoniae antibody titers are significantly associated with acute stroke and transient cerebral ischemia: the West Birmingham Stroke Project. Stroke. 1998; 29: 404–410.
Fagerberg B, Gnarpe J, Gnarpe H, Agewall S, Wikstrand J. Chlamydia pneumoniae but not cytomegalovirus antibodies are associated with future risk of stroke and cardiovascular disease: a prospective study in middle-aged to elderly men with treated hypertension. Stroke. 1999; 30: 299–305.
Glader CA, Stegmayr B, Boman J, Stenlund H, Weinehall L, Hallmans G, Dahlen GH. Chlamydia pneumoniae antibodies and high lipoprotein(a) levels do not predict ischemic cerebral infarctions: results from a nested case-control study in northern Sweden. Stroke. 1999; 30: 2013–2018.
Elkind MS, Lin IF, Grayston JT, Sacco RL. Chlamydia pneumoniae and the risk of first ischemic stroke: the Northern Manhattan Stroke Study. Stroke. 2000; 31: 1521–1525.
Markus HS, Mendall MA. Helicobacter pylori infection: a risk factor for ischaemic cerebrovascular disease and carotid atheroma. J Neurol Neurosurg Psychiatry. 1998; 64: 104–107.
Kolominsky-Rabas PL, Sarti C, Heuschmann PU, Graf C, Siemonsen S, Neundoerfer B, Katalinic A, Lang E, Gassmann KG, von Stockert TR. A prospective community-based study of stroke in Germany: the Erlangen Stroke Project (ESPro): incidence and case fatality at 1, 3, and 12 months. Stroke. 1998; 29: 2501–2506.
Wolfe CD, Tilling K, Beech R, Rudd AG, for the European BIOMED Study of Stroke Care Group. Variations in case fatality and dependency from stroke in western and central Europe. Stroke. 1999; 30: 350–356.
Ridker PM, Buring JE, Shih J, Matias M, Hennekens CH. Prospective study of C-reactive protein and the risk of future cardiovascular events among apparently healthy women. Circulation. 1998; 98: 731–733.
Macy EM, Hayes TE, Tracy RP. Variability in the measurement of C-reactive protein in healthy subjects: implications for reference intervals and epidemiological applications. Clin Chem. 1997; 43: 52–58.
Nieto FJ, Folsom AR, Sorlie PD, Grayston JT, Wang SP, Chambless LE. Chlamydia pneumoniae infection and incident coronary heart disease: the Atherosclerosis Risk in Communities Study. Am J Epidemiol. 1999; 150: 149–156.
Fisher CM. Lacunar strokes and infarcts: a review. Neurology. 1982; 32: 871–876.
Schmal M, Marini C, Carolei A, Di Napoli M, Kessels F, Lodder J. Different vascular risk factor profiles among cortical infarcts, small deep infarcts, and primary intracerebral haemorrhage point to different types of underlying vasculopathy: a study from the L’Aquila Stroke Registry. Cerebrovasc Dis. 1998; 8: 14–19.
Patel P, Mendall MA, Carrington D, Strachan DP, Leatham E, Molineaux N, Levy J, Blakeston C, Seymour CA, Camm AJ, et al. Association of Helicobacter pylori and Chlamydia pneumoniae infections with coronary heart disease and cardiovascular risk factors. BMJ. 1995; 311: 711–714.
Salomaa V, Niemela M, Miettinen H, Ketonen M, Immonen-Raiha P, Koskinen S, Mahonen M, Lehto S, Vuorenmaa T, Palomaki P, Mustaniemi H, Kaarsalo E, Arstila M, Torppa J, Kuulasmaa K, Puska P, Pyorala K, Tuomilehto J. Relationship of socioeconomic status to the incidence and prehospital, 28-day, and 1-year mortality rates of acute coronary events in the FINMONICA myocardial infarction register study. Circulation. 2000; 101: 1913–1918.
Danesh J, Whincup P, Walker M, Lennon L, Thomson A, Appleby P, Wong Y, Bernardes-Silva M, Ward M. Chlamydia pneumoniae IgG titres and coronary heart disease: prospective study and meta-analysis. BMJ. 2000; 321: 208–213.
West DW, Schuman KL, Lyon JL, Robison LM, Allred R. Differences in risk estimations from a hospital and a population-based case-control study. Int J Epidemiol. 1984; 13: 235–239.
Tompkins LS, Schachter J, Boman J, Dowell S, Gaydos CA, Levison ME, Maass M, Madico G, Orfila J, Ouchi K, Peeling RW, Taylor-Robinson D, Stamm WE, Wang SP, Blasi F, Relman D. Collaborative multidisciplinary workshop report: detection, culture, serology, and antimicrobial susceptibility testing of Chlamydia pneumoniae. J Infect Dis. 2000; 181 (suppl 3): S460–S461.
Karvonen M, Tuomilehto J, Pitkaniemi J, Naukkarinen A, Saikku P. Importance of smoking for Chlamydia pneumoniae seropositivity. Int J Epidemiol. 1994; 23: 1315–1321.
Maass M, Gieffers J, Krause E, Engel PM, Bartels C, Solbach W. Poor correlation between microimmunofluorescence serology and polymerase chain reaction for detection of vascular Chlamydia pneumoniae infection in coronary artery disease patients. Med Microbiol Immunol. 1998; 187: 103–106.
Gieffers J, Fullgraf H, Jahn J, Klinger M, Dalhoff K, Katus HA, Solbach W, Maass M. Chlamydia pneumoniae infection in circulating human monocytes is refractory to antibiotic treatment. Circulation. 2001; 103: 351–356.
Danesh J, Wong Y, Ward M, Muir J. Chronic infection with Helicobacter pylori, Chlamydia pneumoniae, or cytomegalovirus: population based study of coronary heart disease. Heart. 1999; 81: 245–247.
Wald NJ, Law MR, Morris JK, Zhou X, Wong Y, Ward ME. Chlamydia pneumoniae infection and mortality from ischaemic heart disease: large prospective study. BMJ. 2000; 321: 204–207.